release: 0.49.0

[blooop]

Ships the agent subprocess contract work: #621, #623 and #622.

docs/agents-using-dl.md publishes four things a caller may write code against. Two of them were not true.

stdout was the command's, except when it wasn't. A dl <ws> -- <command> against a stopped workspace put devpod's whole start transcript on stdout ahead of the command's own output, 26,006 bytes in front of an 18-byte answer, because devpod's logger sends info to stdout and the watched up echoed each line back to the stream it arrived on. Every test of that clause asked a warm workspace, which runs no up at all.

stderr was never the command's. It came back through devpod's stream logger: timestamped, level-tagged, ANSI-coloured, stream_logger.go:492 appended. The page carried a whole section admitting it, beside a 2>&1 workaround. The fix is devpod's own --log-output json on the devpod ssh --command arm, which keeps the level as a field. Its neighbour raw would have fixed this clause and broken the first one: dl recovers a remote exit status from a line it anchors on the word fatal, and raw prints no tag, so dl ws -- 'exit 42' would have quietly stopped exiting 42.

Verified together on a real cold start, streams split:

exit=42                        status recovery survives the transport change
stdout  10 bytes  TO_STDOUT    was 26,006 bytes of build log first
stderr  tail      TO_STDERR    was: 20:27:24 info TO_STDERR stream_logger.go:492

AGENTS.md gained a short section on driving dl from a script, pointing at the contract page rather than restating it.

Release mechanics

Four copies of the version move together, as publish.yml and conda-publish.yml both read [workspace.package] version from rust/Cargo.toml: the manifest, Cargo.lock, and the README's conda badge and dl --version transcript. v0.49.0 has never been tagged (scripts/version_untaken.py agrees), and scripts/changelog_frozen.py confirms all 87 released sections are untouched.

🤖 Generated with Claude Code

Summary by Sourcery

Release version 0.49.0 with reliable subprocess stream handling and documented scripting behavior.

Bug Fixes:

• Keep cold-workspace command output isolated so devpod startup logs do not leak into stdout.
• Preserve command stderr and remote exit-status reporting while maintaining clean stdout/stderr streams.

Enhancements:

• Document the supported subprocess contract for scripts and agents.

Build:

• Bump the workspace and lockfile version to 0.49.0.

Documentation:

• Update release documentation, changelog, version examples, and package badges for 0.49.0.

[sourcery-ai]

Sorry @blooop, you've used your own review budget of 250,000 diff characters for the last 7 days.

You can request another review in 11 hours and 28 minutes by commenting @sourcery-ai review. Upgrade to get a review now.

[sourcery-ai]

Reviewer's guide (collapsed on small PRs)

Reviewer's Guide

This release fixes the agent subprocess transport by using devpod JSON logging on remote command execution, keeping startup/build output off stdout, preserving stderr and remote exit-status semantics, and documenting the verified caller contract. It also synchronizes the 0.49.0 version across release metadata and user-facing documentation.

Sequence diagram for agent subprocess stream separation

sequenceDiagram
    participant Caller
    participant dl
    participant devpod
    participant Command

    Caller->>dl: dl <ws> -- <command>
    dl->>devpod: ssh --command <command> --log-output json
    devpod->>Command: Execute remote command
    Command-->>devpod: stdout and stderr
    devpod-->>dl: JSON log records with level fields
    dl-->>Caller: stdout: command output
    dl-->>Caller: stderr: command errors and startup logs
    dl-->>Caller: remote exit status

Loading

Flow diagram for cold workspace command output

flowchart LR
    Caller[Caller script] --> DL[dl <ws> -- <command>]
    DL --> Start[Start stopped workspace]
    Start --> JSON[devpod JSON log transport]
    JSON --> Streams[Separate stdout and stderr]
    Streams --> Status[Preserve remote exit status]
    Streams --> Caller

Loading

File-Level Changes

Change	Details	Files
Correct the subprocess stream contract for cold-workspace command execution.	Configure the remote SSH command arm to emit JSON logs so devpod startup/build logs do not contaminate stdout and command stderr remains separated. Preserve remote exit-status recovery by continuing to anchor fatal transport records on their structured level/tag. Add documentation and scripting guidance for the four caller-facing agent subprocess guarantees. Record regression coverage for cold starts, split stdout/stderr, and nonzero command exits.	docs/agents-using-dl.md AGENTS.md (tests covering agent subprocess behavior; paths not provided)
Prepare the 0.49.0 release metadata and changelog.	Move the workspace package version to 0.49.0, with the lockfile updated accordingly. Update README Conda badge and --version example. Add the release section while preserving prior changelog history.	rust/Cargo.toml rust/Cargo.lock README.md CHANGELOG.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.